Validation of artificial neural network model for share price

The study’s objective is to justify the use of the ANN for the short-term prediction of share prices, particularly in the banking sector. The assumption is that financial share time-series contain significant non-linearity and that the ANN can be utilized effectively. The ANN model is compared with a linear regression model. Non-linearity is shown by deduction via a comparison of experimental results using the ANN and linear regression models. Furthermore the ANN model is compared with another nonlinear type of model i.e. bilinear model as well. The experiments are based on actual monthly (four-week) period datasets, and the performance of the models is formally evaluated. The conclusions are positive and do merit further experimentation.